Machine Learning is today in the hands of every Sitecore 9.1 developer. No need to have a master in Data science to start. Create your own custom machine learning model and run on Sitecore Cortex in 12 minutes.
In this video I demo how easy it is to create a Machine learning model with Visual Studio ML.NET Model Builder and run this in Sitecore Cortex. The Source Code is on GitHub.
For a quick start I take the Tacoflix-demo from Jason Wilkerson from GitHub, perhaps you've seen it on SUGCON europe.
This demo is about movie recommendation and personalize movie ratings. The code in this demo is just to show how easy it is to work with ML.Net Model builder and not about getting a really good personalize movie recommendation or rating. If you are looking for an easy and good way, take a look at https://aws.amazon.com/personalize/
Personalizing user experience is proven to increase discoverability, engagement, user satisfaction, and revenue, 30% of page views on Amazon are from recommendations. Thinks that work are for example rule-based, if it is almost Christmas you show Christmas movies. Popularity show the current popular movies you can fine-tune it on region, or time of the day. Also Similarity works, for example if your latest rental was the movie Die Hard, you properly also interested in Die Hard 2.
Microsoft Machine Learning, ML Tools
Machine Learning Server (VM, HDInsight Cloudcluster) Create and train models with Python and R. (advanced user) Sitecore Cortex support
Machine Learning Studio (Azure SaaS) Create and train models with drag and drop. (everyone) You can deploy your model as a web services and use that, SaaS you can’t download and not possible to export to ML server.
ML.NET Model Builder The tool from this Video. Create and train models with a view clicks in Visual Studio (developers) Generates C# code, runs everywhere
*All tool are also usable for ML n00bs
Recap Sitecore Cortex™ processing is to Sitecore processing as xConnect is to xDB, it is just a task scheduler build for scale and special for Sitecore and xDB.
Machine Learning can be very simple today, you can start with little effort.
